Re: [Qemu-devel] Runtime-modified DIMMs and live migration issue

[Andrey Korolyov]

>
> Please find the full cli args and two guest logs for DIMM
> initalization attached. As you can see, the freshly populated DIMMs
> are probably misplaced in SRAT ('already populated' messages), despite
> the fact that the initialized ranges are looking correct at a glance.
> When VM is migrated to the destination (with equal RAM device
> configuration) which is simular to a VM with 16G RAM this misplacement
> is causing the mentioned panic in the guest. This should be very
> easily reproducible and the issue is very critical as well, I don`t
> even understand why I missed this issue earlier.

Answering back to myself - I made a wrong statement before, the
physical mapping *are* different with different cases, of course!
Therefore, the issue looks much simpler and I`d have a patch over a
couple of days if nobody fix this earlier.

[  102.757100] init_memory_mapping: [mem 0x240000000-0x25fffffff]
[  102.794016]  [mem 0x240000000-0x25fffffff] page 2M
[  102.798456]  [ffffea0007c00000-ffffea0007ffffff] PMD ->
[ffff88015e400000-ffff88015e7fffff] on node 0
[  102.801853]  [ffffea0008000000-ffffea00081fffff] PMD ->
[ffff880019600000-ffff8800197fffff] on node 0

vs

[    0.411285] init_memory_mapping: [mem 0x240000000-0x25fffffff]
[    0.411288]  [mem 0x240000000-0x25fffffff] page 2M
[    0.416637]  [ffffea0007c00000-ffffea0007ffffff] PMD ->
[ffff880019000000-ffff8800193fffff] on node 0
[    0.422727]  [ffffea0008000000-ffffea00083fffff] PMD ->
[ffff880018c00000-ffff880018ffffff] on node 0